• Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

SMPS CFL inverter heating

Not open for further replies.
Hi all. I am building a mini cfl inverter using SG3525.
The selected frequency is 50 KHz. And I have winded the transformer. I found it in an old inverter. So I don't have its parameters but I could measure Ac wiz 0.75 cm^2.
So Ac= 0.75, frq= 50KHz, Bmax= 1500(?) Vin= 10.5V Vout= 220V
So N(Pri)= 6 and N(sec)= 128
So my problem is I get 250 V at OP when no load and current drawn is .33A. When I connect a cfl (20W) It glows for like half a sec and then suddenly output goes 0 V and current remains 1.15 amps. Sometimes CFL glows 'forever' (Vtg falls to 220V and Mosfets get heated). And then also current is 1.15 amps. But most of the times this doesn't happen. And everything gets heated.
Where is the problem??
I have winded transformer for the first time. Is there any mistake I may have made?
I am using P55 mos.
Plz help me.


Well-Known Member
Most Helpful Member
Can you post your schematic?


Well-Known Member
for a CFL driver, you should use a proper CFL driver chip. There are lots of them, eg IRF do them...it drives the CFL properly. SG3525 is not good for that, you need square waves with variable frequency to regulate the current.
Hello. I think I have found the problem. It was my mistake actually. The problem is in the feedback loop. I used 0.25 W resisters which get heated and the feedback resistance changes and then output falls to zero volts. I need to change the resisters to a higher value. Thanks for your help.

dr pepper

Well-Known Member
Most Helpful Member
Hi Pravin,

As well I think 1500 gauss is a overestimate for a salvaged core, 1000 would be a better assumption, your core could be saturating, maybe pulling the supplybelow the operating point of the sg3525 causing the supply to shut down.

I suggest you build a simple test jig to measure the core saturation pointso you know the core is ok, I have a link to one somewhere.
Thanks for your reply. Some sites suggest 1000 to 1800 Bmax. Thus I have assumed 1500 to be safe.
And how to measure core saturation points?
Also the circuit is drawing 0.3A at 12v from battery with no load and 2.63 amps for 20W CFL. Is this ok??


Well-Known Member
Most Helpful Member
Your circuit looks like a generic voltage-output power supply, and does not have a ballast or any other kind of current limiting necessary for a fluorescent lamp. Also, 1N400x rectifiers are not designed for 50 kHz operation.

Not open for further replies.

Latest threads

EE World Online Articles